AVS › AVS Forum › Audio › Receivers, Amps, and Processors › The Audyssey Pro Installer Kit Thread (FAQ in post #1)
New Posts  All Forums:Forum Nav:

The Audyssey Pro Installer Kit Thread (FAQ in post #1) - Page 72

post #2131 of 4227
4/26/12 at 11:24am
Quote:
Originally Posted by kbarnes701 View Post

Thanks again. Any suggestions for potential improvements?

there's always suggestions, but you said you were done measuring and tweaking...

Quote:
Originally Posted by tandy1000rl View Post

... the after graph succeeded in alerting us that Audyssey could not fully calibrate his system and also indicated where the calibration had failed (i.e. at the crossover frequency). That's a pretty useful function in my eyes, especially ...

yup... the breakdown in communication here is "how useful is the graph?"...

it's very useful in this case to tell us that there is an issue that audyssey is not correcting for... it's usefulness for this task is enhanced by the fact that it makes it immediately obvious that there is an uncorrected issue, EVERYONE noticed it right away...

it's not useful at all when it comes to telling us what is causing the issue or how to correct for it... that's where "we" started poor keith down the path of understanding room modes, measuring repeatedly, and tweaking things for the last several days...

iow, we can trust the audyssey predicted graph to point out the problem, but we can't expect it to do the hard work for us... kinda like the boss...

Quote:
Originally Posted by tandy1000rl View Post

Yes, your graphs very clearly show this effect. Thanks for posting!

you are welcome... lol, yea, in this case, a picture really is worth a thousand words...
Reply
Reply

AVS Top Picks

post #2132 of 4227
4/26/12 at 11:26am
Quote:
Originally Posted by tandy1000rl View Post

Hopefully your subs are both fine, and it's just room interactions and placement driving the need for one to be out of phase with the other, but other possibilities include the LFE port on the AVR being wired incorrectly, or the RCA/XLR cable itself being wired incorrectly. Very very unlikely, and no idea how to test for that, but just pointing out it may not actually be the subs.

Good points. I could test the cable by replacement. And the AVR sub out by swapping subs over I guess. But is it worth it? If I've 'fixed' it by moving the polarity switch all the way to 180 on the 'offending' sub, then does it matter?

Just had another thought - I never tried leaving Sub 2 phase at 0 and moving Sub 1 phase to 180. I wonder if that would show any difference/improvement?
Reply
Reply
post #2133 of 4227
4/26/12 at 11:27am
Quote:
Originally Posted by kbarnes701 View Post

I have considerable sympathy with these points of view

i figured we'd be on the same page there...

Quote:
Originally Posted by kbarnes701 View Post

I'm sure I had more than 4 recommendations... are you using the latest version of the s/w?

i'll check... i suppose it's entirely possible i'm not, although i only downloaded it recently... i only get large, followed by some combination of 40, 60 and 80... edit: yup, 3.5...
Reply
Reply
post #2134 of 4227
4/26/12 at 11:31am
Quote:
Originally Posted by kbarnes701 View Post

My speaker crossover points are 3kHz between the tweeter and the midrange and 300 Hz between the midrange and the woofer. How does that fit in with the mid-range compensation thing?

Likely your speakers are benefiting from the midcomp.
Reply
Reply
post #2135 of 4227
4/26/12 at 11:33am
Quote:
Originally Posted by kbarnes701 View Post

I'm sure I had more than 4 recommendations... are you using the latest version of the s/w?

It is probably dependent on the measurements. If there are five that are deemed acceptable, you get five. If only four, then ...

Jeff
Reply
Reply
post #2136 of 4227
4/26/12 at 11:37am
Quote:
Originally Posted by tandy1000rl View Post


No derision intended here, but the after graph succeeded in alerting us that Audyssey could not fully calibrate his system and also indicated where the calibration had failed (i.e. at the crossover frequency). That's a pretty useful function in my eyes, especially for the intended audience of this technology, professional Audyssey installers who aren't likely to spend weeks or months on a single calibration job like we as enthusiasts may.

Also, the fix Keith put into place (flip sub 2 out of phase) was born out of his own knowledge and experience--that issues around the crossover frequency are driven by summation/cancellation issues due to phase--so that's where he began experimenting. His other measuring system didn't reveal any insights there, other than greater graph resolution. Point is, he could have done the same experimentation with the Audyssey certificate graphs only and arrived at the same place--a successful calibration. ;-)

That said, what insights did you glean from his independent measurements that weren't already in his Apr 20 certificate?

Honestly, this issue is old. If you feel Audyssey graphs work for you to tune your system then go for it. And I mean that sincerely, no animus. The last time I checked room modes occurred at different positions in a room, yet Audyssey sub graphs display a single FR representative of all listening positions. Perhaps Audyssey is able to pull off this feat contrary to the physics as I understand them. It has been pointed out that I'm no expert on this topic, and I don't disagree with that, nor have I purported to be one.

However, I do have measurements and eyeballs that have conspired to make me stubbornly hold on to the admittedly uneducated concept that a single averaged graph below Schroeder *might* not be accurate for every listening position in a room.
Reply
Reply
post #2137 of 4227
4/26/12 at 11:39am
Quote:
Originally Posted by tandy1000rl View Post

No derision intended here, but the after graph succeeded in alerting us that Audyssey could not fully calibrate his system and also indicated where the calibration had failed (i.e. at the crossover frequency). That's a pretty useful function in my eyes, especially for the intended audience of this technology, professional Audyssey installers who aren't likely to spend weeks or months on a single calibration job like we as enthusiasts may.

Yes, it was the disappointment of seeing that serious looking dip in the sub 'after' graph that started the ball rolling and inspired me to dig out the OmniMic and investigate further. In that sense alone, the after graph was useful to me.

I take Gooddoc's point that the after graph doesn't tell me much about what is going on at the MLP, but then it isn't meant to is it?

Quote:
Originally Posted by tandy1000rl View Post

Also, the fix Keith put into place (flip sub 2 out of phase) was born out of his own knowledge and experience--that issues around the crossover frequency are driven by summation/cancellation issues due to phase--so that's where he began experimenting. His other measuring system didn't reveal any insights there, other than greater graph resolution. Point is, he could have done the same experimentation with the Audyssey certificate graphs only and arrived at the same place--a successful calibration. ;-)

Yes, I did consider, briefly, just rerunning Pro again with the phase reversed and seeing what I would get. In fact, I now know what I would have got - the result I actually did get, which was a flat sub 'after' graph. I decided to use the OmniMic to fiddle about simply because it is so much quicker than running a Pro calibration from scratch. If hadn't owned the OM I would have still got to the same end point, eventually.
Reply
Reply
post #2138 of 4227
4/26/12 at 11:49am
Is the Curve Editor something I could use to tweak out that dip at about 175 Hz, or is it for something else? I haven't even looked at it yet. Or is the concensus that that dip isn't all that important? Thanks.

Reply
Reply
post #2139 of 4227
4/26/12 at 11:53am
I don't think that dip is correctable electronically.

Jeff
Reply
Reply
post #2140 of 4227
4/26/12 at 12:01pm
Quote:
Originally Posted by kbarnes701 View Post

Is the Curve Editor something I could use to tweak out that dip at about 175 Hz, or is it for something else? I haven't even looked at it yet. Or is the concensus that that dip isn't all that important? Thanks.


I assume that graph is with just your LR and subs?

If so, that is below Schroeder but above your subs. Therefore it's caused by your mains. It *may* be possible to shift your mains around a bit to correct that, depending on its axis.


Is it important? Only if you can hear it.
Reply
Reply
post #2141 of 4227
4/26/12 at 12:01pm
@keith...

looks like a null to me... i think you'd just be feeding power into a black hole... i'd have to run the calculator again, but my "guess" is that's a room mode issue... corner traps likely wouldn't "hurt" that...

it's a reasonably deep null, but it's not real wide... it easily falls into the "more than liveable, at least until the next stage of geekdom" category...

as someone else noted, moving the main speakers a bit may help with that, but since you have likely placed the speakers for best staging/imaging, moving them would likely hurt the overall package more than help it...

it can be easy to focus on a seemingly large specific issue, and forget about the "bigger picture"...

fwiw, it's actually closer to 155-160hz... don't forget the graph doesn't have a linear x axis...
Reply
Reply
post #2142 of 4227
4/26/12 at 12:15pm
Quote:
Originally Posted by pepar View Post

I don't think that dip is correctable electronically.

Jeff

Quote:
Originally Posted by Gooddoc View Post

If so, that is below Schroeder but above your subs. Therefore it's caused by your mains. It *may* be possible to shift your mains around a bit to correct that, depending on its axis.


Is it important? Only if you can hear it.

Quote:
Originally Posted by ccotenj View Post

@keith...

looks like a null to me... i think you'd just be feeding power into a black hole... i'd have to run the calculator again, but my "guess" is that's a room mode issue... corner traps likely wouldn't "hurt" that...

it's a reasonably deep null, but it's not real wide... it easily falls into the "more than liveable, at least until the next stage of geekdom" category...

as someone else noted, moving the main speakers a bit may help with that, but since you have likely placed the speakers for best staging/imaging, moving them would likely hurt the overall package more than help it...

it can be easy to focus on a seemingly large specific issue, and forget about the "bigger picture"...

fwiw, it's actually closer to 155-160hz... don't forget the graph doesn't have a linear x axis...

Thanks guys. I am not hearing anything odd, so I think I will sit back, get a cold one from the fridge and watch some movies!
Reply
Reply
post #2143 of 4227
4/26/12 at 12:30pm
Great idea! I think you slayed the elephant in the room. You can live with the mice for a while...
Reply
Reply
post #2144 of 4227
4/26/12 at 12:33pm
Quote:
Originally Posted by Gooddoc View Post

Great idea! I think you slayed the elephant in the room. You can live with the mice for a while...

Haha - nicely put.
Reply
Reply
post #2145 of 4227
4/26/12 at 12:37pm
Quote:
Originally Posted by kbarnes701 View Post

Thanks guys. I am not hearing anything odd, so I think I will sit back, get a cold one from the fridge and watch some movies!

Music (jazz) with a prominent bass guitar would be the easiest content with which to hear something like this. Not possible with cinema content.
Reply
Reply
post #2146 of 4227
4/26/12 at 1:25pm
Quote:
Originally Posted by pepar View Post

Music (jazz) with a prominent bass guitar would be the easiest content with which to hear something like this. Not possible with cinema content.

Stanley Clarke always helps me out Although I only use my HT system for movies, I always check my SQ using the kind of music you suggest. I know what music sounds like for one thing, whereas I have little idea of what a fighting junkyard robot should sound like, or a man with a big round sparkly thing embedded in his chest escaping from a cave, or a man in a black suit with pointy ears inside the coolest car *ever*...

This time around, Stan and I could hear nothing amiss
Reply
Reply
post #2147 of 4227
4/26/12 at 1:34pm
Quote:
Originally Posted by kbarnes701 View Post


Stanley Clarke always helps me out Although I only use my HT system for movies, I always check my SQ using the kind of music you suggest. I know what music sounds like for one thing, whereas I have little idea of what a fighting junkyard robot should sound like, or a man with a big round sparkly thing embedded in his chest escaping from a cave, or a man in a black suit with pointy ears inside the coolest car *ever*...

This time around, Stan and I could hear nothing amiss

Good choice. You just inspired me to listen to some RTF - never fails to put a smile on my face
Reply
Reply
post #2148 of 4227
4/26/12 at 1:54pm
Quote:
Originally Posted by kbarnes701 View Post


This time around, Stan and I could hear nothing amiss

Then settle in with your favorite beverage and enjoy!
Reply
Reply
post #2149 of 4227
4/26/12 at 2:20pm
Quote:
Originally Posted by ccotenj View Post

a few "thoughts"...

money seat looks pretty good... a bit messy from 95-350hz, pretty much as expected... nice and flat above 500hz, even with 1/24th smoothing... aud didn't try to do much there, again as expected (and desired)...

Hey Chris,

Looking at your money seat graphs, you show 10-200 Hz and 10-500 Hz graphs with Audyssey on. The graphs look quite different below 40 Hz. Assuming the mic wasn't moved, shouldn't the graphs be identical?
Reply
Reply
post #2150 of 4227
4/26/12 at 2:29pm
@jerry...

actually, i'm looking at them again, and they look similar down to about 15hz... the different frequency response scale is what makes them appear different on the surface, unless i'm missing something (which is entirely possible )...

below 15hz, i'm writing off as microphone aberration, etc. at this point...
Reply
Reply
post #2151 of 4227
4/26/12 at 2:43pm
If you're looking for some music to test your system, here's a link to some free downloads from Ken Kreisel:

http://www.kkprofessionalsound.com/downloads.htm

I'm told by Ken that Flamenco Fever, Hot Stix, and Encore: Set Down Servant are all very good for bass transient response tests.
Reply
Reply
post #2152 of 4227
4/26/12 at 3:13pm
Quote:
Originally Posted by kbarnes701 View Post

My speaker crossover points are 3kHz between the tweeter and the midrange and 300 Hz between the midrange and the woofer. How does that fit in with the mid-range compensation thing?



..

I do not think the midrange compensation dip really has anything to do with the crossover. I just think you should try it with/without the dip at some point to see which you prefer subjectively. I would use music rather than movies in evaluating it. Pick some of your best recordings. Then go with what your gut tells you sounds best.
Reply
Reply
post #2153 of 4227
4/26/12 at 3:17pm
omg i remember that flamenco fever title. ken was a pioneer. iirc, some of his recordings went direct-to-master ...as portrayed in that pic.
Reply
Reply
post #2154 of 4227
4/26/12 at 3:20pm
iirc, crossover is mentioned in audyssey's spiel about midcomp, but i think it goes mainly to changing dispersion of drivers around that region.
Reply
Reply
post #2155 of 4227
4/26/12 at 3:34pm
Quote:
Originally Posted by pepar View Post

iirc, crossover is mentioned in audyssey's spiel about midcomp, but i think it goes mainly to changing dispersion of drivers around that region.

The argument for mid bass compensation has to do with better dialogue due to problems at the mid to treble crossover. This is discussed by Chris in an ask audyssey section. Google and you will find. B&W introduced this as an option in their 801 matrix circa 1980. Not sure one size fits all makes sense. Clearly not useful for speakers with no xover in the mid bass to treble region. Search mid range compensation 21 Jan.2011. Ask audyssey.
Reply
Reply
post #2156 of 4227
4/26/12 at 4:13pm
Quote:
Originally Posted by Gooddoc View Post


However, I do have measurements and eyeballs that have conspired to make me stubbornly hold on to the admittedly uneducated concept that a single averaged graph below Schroeder *might* not be accurate for every listening position in a room.

You are absolutely correct that an averaged graph is not "correct" for every listening position in the room, in the sense that it does not portray the actual response at any one position. But, Audyssey is designed from the ground up to work with a spatial average, as is common, accepted acoustical practice. It identifies departures from the target curve which are persistent and confirmed at multiple positions, and it sets EQ filters accordingly. This avoids overreacting to the localized response at one point, which per the REG piece I cited, is known to be "spatially unstable", i.e., it is not usually like response at other positions for the same speaker channel.

In any case, the Audyssey graphs portray what Audyssey is designed to do: determine response at multiple positions for each channel and determine one single set of averaged filters across the frequency and time domains for each channel. The graphs merely portray the results of this before and after.

I do not honestly know what you expect Audyssey to do with the graphs. I get the impression that you want it to provide graphs of before/after response at each mike position. Let's see. With a 7.1 or 7.2 system, we get 16 graphs now. If we did 8 positions, this scheme would provide 128 graphs. I do not know about you, but I think that would be extremely unwieldy and difficult to interpret and it is not consistent with how Audyssey works anyway. It is not geared as a measurement and display tool. It is an EQ tool. Its graphic results are merely a byproduct of that.

But, you already have the much simpler measurement and display tool for independent measurements. So, why not just use that to check before/after response for your purposes and call it a day.
Reply
Reply
post #2157 of 4227
4/26/12 at 4:18pm
Quote:
Originally Posted by sdrucker View Post


One thing--how many iterations, and by what distance increments, did you explore the distance range? After all, varying each sub individually would result in 144 iterations--and that's just with 0.1 increments and only moving a foot from the distances generated by Pro...it's not like you can predict the bass response as distance varies from an algorithm...or is there?

Stuart

I thought you'd never ask.

When you change the sub distance relative to the nearby satellite distance, you are simply changing the angle between two sinewaves so they sum to a different value. I.e., if you move one relative to the other by one half wavelength (180 degrees), you have two out-of-phase sinewaves and a pretty deep dip (infinitely deep if you count only the two equal amplitude direct waves). Move one relative to the other by less than one half wavelength and you have a less deep dip. Make them in phase and you have a flat curve (if the two have the same SPL).

Below is a table for an 80 Hz crossover showing an estimate of how much to change the sub to correct a given dip at the crossover. The phase correction you will be making is also shown. This is based on the direct waves so may not be very precise in practice.

Notice that the sub distance can be off by a foot or more with very little impact. Hence most distance corrections will be more than a foot.

EDIT: .... found a bug in my spreadsheet. I wasn't calculating the peak of the combined sub and satellite sinewave correctly. I think the updated table is correct.

While I believe this table is correct, note that there can be considerable error with very little effect. Plus or minus 50 degrees (about 2 feet) keeps the dip to less than 1 db. Hence the distance correction in the table should easily put you in the ball park.



If anyone would like the Excel sheet wherein I derived this table, pm me.
Reply
Reply
post #2158 of 4227
4/26/12 at 5:01pm
Quote:
Originally Posted by fitzcaraldo215 View Post


You are absolutely correct that an averaged graph is not "correct" for every listening position in the room, in the sense that it does not portray the actual response at any one position. But, Audyssey is designed from the ground up to work with a spatial average, as is common, accepted acoustical practice. It identifies departures from the target curve which are persistent and confirmed at multiple positions, and it sets EQ filters accordingly. This avoids overreacting to the localized response at one point, which per the REG piece I cited, is known to be "spatially unstable", i.e., it is not usually like response at other positions for the same speaker channel.

In any case, the Audyssey graphs portray what Audyssey is designed to do: determine response at multiple positions for each channel and determine one single set of averaged filters across the frequency and time domains for each channel. The graphs merely portray the results of this before and after.

I do not honestly know what you expect Audyssey to do with the graphs. I get the impression that you want it to provide graphs of before/after response at each mike position. Let's see. With a 7.1 or 7.2 system, we get 16 graphs now. If we did 8 positions, this scheme would provide 128 graphs. I do not know about you, but I think that would be extremely unwieldy and difficult to interpret and it is not consistent with how Audyssey works anyway. It is not geared as a measurement and display tool. It is an EQ tool. Its graphic results are merely a byproduct of that.

But, you already have the much simpler measurement and display tool for independent measurements. So, why not just use that to check before/after response for your purposes and call it a day.

Whew! I don't know how to reply to a post that has made all the very same points I've made in prior posts(although I admit you said it better ) but to say....I agree.

The only thing you got wrong is that I expect nothing from Audyssey beyond what it does. Not sure where you got that from.

So I'll repeat, the Audyssey after graphs are as useful as tits on a bull. Anyone else want to beat me up for that? .
Reply
Reply
post #2159 of 4227
4/26/12 at 5:03pm
Quote:
Originally Posted by hclarkx View Post

I thought you'd never ask.

When you change the sub distance relative to the nearby satellite distance, you are simply changing the angle between two sinewaves so they sum to a different value. I.e., if you move one relative to the other by one half wavelength (180 degrees), you have two out-of-phase sinewaves and a pretty deep dip (infinitely deep if you count only the two equal amplitude direct waves). Move one relative to the other by less than one half wavelength and you have a less deep dip. Make them in phase and you have a flat curve (if the two have the same SPL).

Below is a table for an 80 Hz crossover showing an estimate of how much to change the sub to correct a given dip at the crossover. The phase correction you will be making is also shown. This is based on the direct waves so may not be very precise in practice.

Notice that the sub distance can be off by a foot or more with very little impact. Hence most distance corrections will be more than a foot.



If anyone would like the Excel sheet wherein I derived this table, pm me.

I had a feeling something like this existed.....I'm not a mathematician in my day job, but since I use tools like Crystal Ball for product optimization in my biz, I understand the concept, even if not how you got there.

You should get a cigar - or one of ccotenj's IPAs LOL......just for thinking of this!

If I understand this correctly, you can use the db adjustment on the left in conjunction with the adjustment in ft. on each sub. The art will be experimenting in the exact mix between the subs, when you have two or more subs (since with multiple subs, they'll not neccessarily have the same effect). My guess is that you'd halve the adjustment as a start, and then go up/down with the other held constant, repeating the exercise for a second sub. but if there's merit in going the full adjustment for one sub, and then see how the other sub's adjustments influence the total, that might work too.

One thing: I have a powered center (Deftech CS-8080HD). Do you have any thoughts about how a different crossover - say 90 Hz - than the mains (which I'm keeping @ 80) might influence the adjustment?


PM is on its way.....I'd love to try this tool when I redo my own measurements (probably next week).

Thanks!
Stuart
Reply
Reply
post #2160 of 4227
4/26/12 at 5:13pm
Quote:
Originally Posted by hclarkx View Post

I thought you'd never ask.

When you change the sub distance relative to the nearby satellite distance, you are simply changing the angle between two sinewaves so they sum to a different value. I.e., if you move one relative to the other by one half wavelength (180 degrees), you have two out-of-phase sinewaves and a pretty deep dip (infinitely deep if you count only the two equal amplitude direct waves). Move one relative to the other by less than one half wavelength and you have a less deep dip. Make them in phase and you have a flat curve (if the two have the same SPL).

Below is a table for an 80 Hz crossover showing an estimate of how much to change the sub to correct a given dip at the crossover. The phase correction you will be making is also shown. This is based on the direct waves so may not be very precise in practice.

Notice that the sub distance can be off by a foot or more with very little impact. Hence most distance corrections will be more than a foot.



If anyone would like the Excel sheet wherein I derived this table, pm me.

So this would be helpful if one's subwoofer does not have a variable phase knob (i.e. you can get the same result by physically moving the sub, changing its delay in the AVR, or dialing in the needed correction using the phase knob)?

Taking that a step further, if there is a dip at the crossover, Audyssey can compensate for it to a point by boosting that frequency with its filters, or if the dip is too deep and falls outside of Audyssey's range, then delaying subwoofer distances and/or varying the phase knob should be action item #1 before attempting to recalibrate. Don't have to get it to perfect integration before calibration, but rather get it within the correctable threshold and Audyssey's filters can take over from there.

If so, do we think the blurb in the setup guides and FAQs about setting the phase to 0 should be footnoted accordingly?

Very interesting table indeed!

Edit: upon further thought, doesn't Audyssey have everything it needs to compute and adjust for this automatically? That is, shouldn't it have dialed in an extra 7 feet of distance in Keith's 2nd subwoofer because it knew it was out of phase with respect to the 1st subwoofer and the mains based on relative arrival times of the test chirps? Isn't that the whole point of subwoofer physical distance not matching subwoofer acoustical distance? /confused/
Reply
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Receivers, Amps, and Processors

AVS Top Picks

AVS › AVS Forum › Audio › Receivers, Amps, and Processors › The Audyssey Pro Installer Kit Thread (FAQ in post #1)